Journal of Experimental Marine Biology and Ecology, L 216 (1997) 171±189 Benthic faunal responses to variations in patch density and patch size of a suspension-feeding bivalve Robert B. Whitlatch* , Anson H. Hines1 , Simon F. Thrush, Judi E. Hewitt, Vonda Cummings National Institute of Water and Atmospheric Research, 100 Aurora Terrace Road, P.O. Box 11-115, Hamilton, New Zealand Abstract Responses of benthic macrofauna and epibenthic predators/sediment disturbers to controlled density (0, 12, 120, 600 and 1200 individuals m222 ) and patch size (0.25, 1.0 and 9.0 m ) manipulations of the venerid bivalve Austrovenus stutchburyi (Gray) were examined on an intertidal sand¯at. Bivalve density manipulations greater than 600 m22 had a mixed in¯uence on macrofauna colonisation; two species were reduced, three enhanced and three were unaffected. There was no clear functional group-related pattern as to which species were affected by Austrovenus. The most pronounced in¯uence of high bivalve densities was the reduction in the abundance and alteration of the size-structure of post-set (ca. 250±360 mm) tellinid bivalves, Macomona liliana Iredale. Patch size manipulations of Austrovenus had no measurable effect on macrofauna colonisation, probably because patch density (120 bivalve m22 ) was too low. While a variety of epibenthic predators were present at the study site, Austrovenus stutchburyi mortality estimates in both experiments were relatively low (0.01±0.03% individuals lost day21 ) and were independent of both density and patch size. In contrast, the proportion of nipped bivalve siphons was relatively high (11±37%) in both experiments. While proportions of nipped siphons were similar across the range of manipulated densities, siphon browsing was more than 2 times higher in 9.0 m22 plots than 0.25 m plots. Scale-dependent foraging of sub-lethal marine benthic predators is previously unreported and our results illustrate the necessity for conducting studies on predator±prey dynamics at ecologically meaningful spatial scales for both the predator and prey species. 1997 Elsevier Science B.V. Keywords: Sand¯at; Colonisation; Adult±juvenile interaction; Patch dynamics *Corresponding author. Present address: Department of Marine Sciences, University of Connecticut, Groton, CT 06340, USA. 1Present Address: Smithsonian Environmental Research Center, P.O. Box 28, Edgewater, Maryland 21037, USA. 0022-0981/97/$17.00 1997 Elsevier Science B.V. All rights reserved. PII S0022-0981(97)00095-6 172 R.B. Whitlatch et al. / J. Exp. Mar. Biol. Ecol. 216 (1997) 171 ±189 1. Introduction Considerable attention has been directed to studying the role resident adult infaunal invertebrates have determining the structure of assemblages in soft sediment habitats. Apart from potential taxonomic biases and density-dependent variations in adult± juvenile interactions, Olafsson et al. (1994) recently concluded that deposit-feeding and predatory infaunal invertebrates typically act to reduce recruitment, whereas suspension- feeding molluscs generally exert much less effect on the recruitment dynamics of infaunal organisms. Others, however, have shown that suspension-feeders can have strong interactions of benthic±pelagic coupling in shallow water (Dame, 1993, for a recent review) and may affect infaunal community structure by removing small planktonic larvae before they reach the sediment surface (e.g., Kristensen, 1957; Woodin, 1976; Andre et al., 1993). In addition, modi®cation of local boundary ¯ow patterns by physical structures (e.g., Eckman, 1985) and siphonal currents (e.g., Monismith et al., 1990; O'Riordan et al., 1993) may also in¯uence larval settlement. Suspension-feeding bivalves also may interact with deposit-feeding populations by removing phytoplankton and/or by deposition of faeces or pseudofaeces (e.g., Cloern, 1982; Of®cer et al., 1982; Lin and Hines, 1994). Large infaunal bivalves that live near the sediment surface can also affect other species by providing refugia from predation (e.g., Luckenbach, 1987; Peterson and Black, 1993). Lastly, several authors have noted the spatial scale at which many experimental studies of benthic suspension feeders have been conducted may limit the generality of conclusions as well as explain the apparent inconsistencies of the interactions of resident suspension-feeding bivalves and other benthic invertebrates (e.g., Black and Peterson, 1988; Andre and Rosenberg, 1991; Andre et al., 1993). Since most benthic organisms are patchily distributed over a range of spatial scales (e.g., Thrush, 1991), it is important to study the major variables characterising these spatial arrangements: organism density and three components of scale (grain or patch size, lag or distance between patches and extent or area occupied (Thrush et al., 1997a). These components of the spatial arrangement of density can be independently manipu- lated in experiments, at least up to the 10's m scale, to identify their role in affecting ecological processes. The goal of this study was to assess benthic fauna responses to manipulations of patch size (grain) and density of a common suspension-feeding bivalve, Austrovenus stutchburyi (Gray) inhabiting an intertidal sand¯at habitat. The focus of our study was to assess how macroinfaunal recruitment and epibenthic bivalve predators/sediment disturbers responded to controlled manipulations of Au- strovenus stutchburyi density and patch size. Experimental density manipulations were restricted to naturally meaningful levels (e.g., Olafsson et al., 1994; Thrush et al., 1996) and similarly, variations in spatial arrangement re¯ected natural patterns of the shell®sh at the study site. For example, Hewitt et al. (1996) found an average Austrovenus density (mean6SD) of 7406458 m222 in a 50 m area of study site with high bivalve densities. They also found spatial structure of Austrovenus at 1.3 and 3 to 3.7 m and also indicated smaller-scale variations at about 30 cm. Our experiments were designed to test several hypotheses. Those related to the effects of altering Austrovenus stutchburyi density were: (a) macrofauna recruitment will be R.B. Whitlatch et al. / J. Exp. Mar. Biol. Ecol. 216 (1997) 171 ±189 173 reduced at high bivalve densities due to shell®sh ®ltration and space limitation (e.g., Woodin, 1976; Cloern, 1982; Lin and Hines, 1994; Andre et al., 1993), (b) macroinfauna recruitment and/or immigration responses will be dependent upon functional group (e.g., tube-building species will be negatively affected by dense populations of suspension feeders while burrowing species will remain unaffected (Woodin, 1976; Jensen, 1984; Flach, 1992)), (c) at intermediate bivalve densities, macrofauna recruitment will be enhanced due to sediment stabilisation by shell structure and possible enrichment by production of pseudofaeces for surface deposit feeders (e.g., Commito and Boncavage, 1989) and (d) due to predator aggregation foraging behavior, both lethal and sublethal predation will increase with bivalve density (e.g., Hassell and May, 1985; ``prey taxis'': Kareiva and Odell, 1987). Hypotheses related to manipulation of grain were: (a) larger patches of suspension feeders will result in lower macrofauna recruitment for species with low ¯ux rates or small scales of movement due to a reduced ratio of perimeter to area (e.g., Smith and Brumsickle, 1989; Thrush et al., in press) and (b) lethal and sublethal predation on the bivalves should be most pronounced on largest patches of prey due to aggregative predator responses (e.g., Wilson, 1990; Colwell and Landrum, 1993; Horne and Schneider, 1995). These hypotheses were tested by manipulating either patch size with a constant density of Austrovenus or bivalve density in a patch of constant grain. 2. Methods 2.1. Study species Austrovenus stutchburyi is widely distributed throughout the estuaries and harbours of New Zealand (Morton and Miller, 1973) and is often a dominant member of macrofauna communities (Paul, 1966; Cassie and Michael, 1968; Blackwell, 1984; Pridmore et al., 1990; Turner et al., 1995). This venerid bivalve is an obligate suspension feeder with short siphons that restrict it to the upper 1±3 cm of the sediment column. Predators of Austrovenus in the nearshore zone include wading shore birds (South Island Pied Oystercatcher Haematopus ostralegus ®nschii Martens, Red Knot Calidris canutus rogersi (Mathews), Eastern Bar-Tailed Godwit Limosa lapponica baueri Naumann) and gulls (Southern Black-backed Gull Larus dominicanus Lichtenstein, Red-billed Gull Larus novaehollandiae scopulinus Foster), as well as eagle rays (Myliobatis tenuicaudatus (Hector), crabs and gastropods (e.g., Knox, 1980; Thrush et al., 1991, 1994; Cummings et al., 1997). Several species of ¯ounder (Rhombosolea leporina Gunter and R. plebeia Richardson) found at the study site are reported to in¯ict considerable siphon-nipping on Austrovenus (Webb, 1973; Kilner, 1974). 2.2. Study area This study was conducted in the mid-intertidal zone of a sand¯at located adjacent to Wiroa Island in Manukau Harbour, New Zealand (37 8029 S, 174 8419 E). Sediments in this area are composed of predominantly ( . 90%) well-sorted ®ne sand, and the 174 R.B. Whitlatch et al. / J. Exp. Mar. Biol. Ecol. 216 (1997) 171 ±189 sediment surface is relatively homogeneous with the main physical features being sand ripples (0±2 cm high) and small drainage runnels. Biogenic structures common throughout the intertidal zone are primarily feeding pits created by eagle rays, siphon traces of Macomona liliana